DTL Posted June 6, 2007 Share Posted June 6, 2007 Hi All, I posted this somewhere else because I couldn't figure out how to start a new thread... I think I got it now.... I Have a 91 Subaru Lagacy wagon w/ AWD, ABS, automatic transmission, 2.2 litre engine sitting in my driveway waiting for me to fix it. It died 2 years ago the morning after a long trip. My wife said it was running rough towards the end of the trip in local traffic. The next morning, it started for a second, then stalled and has not started since. it has 180k of mostly Highway miles and was running well. The timing chain was replaced a few months before it's demise. I checked compression which seemed good. Turned out that the Ignition transistor was not firing one side of the spark coil. For $50. I got a good used xistor, popped it in and cranked it. No start. The new trasistor got immediately toasted. I have been staring at this Subaru for the last 9 months wondering if it is worth trying to replace the coil and the Xistor (perhaps the computer too?) The Coil reads about an ohm from the center tap to each of the lead ends and infinity to the spark plug leads. I am GUESSING that HV spark is leaking back to the xistor somehow. Any thoughts on this? I live on long Island and would consider selling this vehicle for a few hundred bucks to someone who wants to fix it or use it for a parts car. The ABS system looked shiny as if it was replaced not long before we bought the car and put our 20k miles on it. I have to clear the driveway within a couple of weeks, but, this was such a great car, I hate to sell it for scrap. Link to comment Share on other sites More sharing options...
nipper Posted June 6, 2007 Share Posted June 6, 2007 SOunds like a dead short someplace. The coil resistance seems low too. What are you calling an xistor? nipper Link to comment Share on other sites More sharing options...
ron917 Posted June 6, 2007 Share Posted June 6, 2007 xistor = transistor Used as an electronic switch in the ignition system, it pulses the primary side of the coil. I think the car manufactuerers usually call the device an ignitor. Link to comment Share on other sites More sharing options...
nipper Posted June 6, 2007 Share Posted June 6, 2007 xistor = transistor Used as an electronic switch in the ignition system, it pulses the primary side of the coil. I think the car manufactuerers usually call the device an ignitor. Yup an ignitor. DO you have a manual for the car? Its not a complicated circuit and can be easily traced out. You may have a shorted coil . nipper Link to comment Share on other sites More sharing options...
DTL Posted June 6, 2007 Author Share Posted June 6, 2007 xistor = transistor Used as an electronic switch in the ignition system, it pulses the primary side of the coil. I think the car manufactuerers usually call the device an ignitor. Yes. The computer seems to send logic 0's alternately to the ignitor's 2 inputs, the outputs of which seem to provide a path to ground for the +12v at the coil pack center tap alternately through the 2 primary coils. I will recheck the primary ohms, but my guess is that a HV short might not show up with any ohms test. I thought I checked for bad grounding and found everthing ok I would expect some circuitry like a flyback diode internal to the ignitor to catch any primary HV... I will recheck it all later today. I was hoping for a confirmation of my suspicions from one of you techies who might have come accross this in the past. I put in a parts requst to Quality Auto Parts internet parts locator who claim to have what I need for $68.00 plus shipping. I would prefer to buy from someone local... has any of you folks had any troubles with the parts locators, this one has a 251 area code, I am not sure where that is... Link to comment Share on other sites More sharing options...
Jimmyowensfan20 Posted June 6, 2007 Share Posted June 6, 2007 how much you want for it? Link to comment Share on other sites More sharing options...
DTL Posted June 6, 2007 Author Share Posted June 6, 2007 OOOOOOOOK, I bought a used coil and ignitor from a junkie errrr junkyard salvage place. http://www.main.experiencetherave.com:8080/subaru_manual_scans/FSM_Scans/ignitionsystem.jpg Gives a useful schematic, but, It looks like 2 spark plugs are on the same secondary with chasis ground ( or engine block) as a center tap(????????). This implies to me that one of the spark plugs on each side is a cathode and one is an anode. and that each time the spark has to jump both spark gaps to complete the secondary circut. I can see how this could fry my ignitor if the wiring from the ignitor to the coil gets exposed to the secondary. Could someone confirm this?, before i hook it all up i wanna makesure there will be no more expensive problems. I am going to beef up the insulation between the coil wiring and the motor block, any other suggestions will be appreciated. :burnout: Link to comment Share on other sites More sharing options...
hankosolder2 Posted June 6, 2007 Share Posted June 6, 2007 Ah, another person with an electronics backround. Yes, the plugs paired on one bank of the coil get opposite polarity pulses. It's called a wasted spark system. It seems remotely possible that if you have a open in your ignition secondary circuit (most likely a plug wire) the high voltage could arc internally from the coil pack's secondary winding to the primary winding and fry the transistor in the ignitor. You could check the resistance of the plug wires. I would scrutinize the low voltage part of the circuit first though. Does the coil have equal resistance from the centertap on the primary to each primary leg? Another possiblity might be if the computer is holding one of the outputs to the ignitor constantly low or high...rather than pulsing. The ignitor is probably not designed to dump 100% duty cycle DC into the relatively low resistance of the coil pack. You could check for the proper drive signals going to the ignitor, check the wiring between the ignitor and the coil pack for shorts, check the plug wires for opens, replace the coil pack and ignitor simultaneously and you'll probably be in business. Let us know what you find! Nathan OOOOOOOOK, I bought a used coil and ignitor from a junkie errrr junkyard salvage place. http://www.main.experiencetherave.com:8080/subaru_manual_scans/FSM_Scans/ignitionsystem.jpg Gives a useful schematic, but, It looks like 2 spark plugs are on the same secondary with chasis ground ( or engine block) as a center tap(????????). This implies to me that one of the spark plugs on each side is a cathode and one is an anode. and that each time the spark has to jump both spark gaps to complete the secondary circut. I can see how this could fry my ignitor if the wiring from the ignitor to the coil gets exposed to the secondary. Could someone confirm this?, before i hook it all up i wanna makesure there will be no more expensive problems. I am going to beef up the insulation between the coil wiring and the motor block, any other suggestions will be appreciated. :burnout: Link to comment Share on other sites More sharing options...
DTL Posted June 7, 2007 Author Share Posted June 7, 2007 I have a little electronics background... enough to get me in trouble...lol The ignitor and coil pack new wires are installed.... but somehow dropped a wrench on top of my meter lcd screen which now displays some chineese dialect in broken-alphanumeric. It is getting dark, tomorrow i will find my other meter and bring my digital scope outside and have a look at the control signal waveforms. Any idea as to how long a pulse would be toooo long? Well, only one side of the ignitor burned so, if something is wrong one will be much longer than the other. I was SOOOO tempted to try and start it up. But, I would rather not have to buy both the ECU computer and another ignitor (how much do the junkyards get for these?). BTW the F27 ground lug is where? under dash? the scan is hard to read but looks like it is connecting to terminals 7 & 8 on ECU I hate working under the dash, I practically need to hang upside down to do anything and/or see anything. http://www.main.experiencetherave.com:8080/subaru_manual_scans/FSM_Scans/ignitionsystem.jpg I am guessing that the ignitor heatsink is not part of the circut. Thanks for the reply, I was afraid I would be going this one alone! Bill Link to comment Share on other sites More sharing options...
OB99W Posted June 7, 2007 Share Posted June 7, 2007 [...]The Coil reads about an ohm from the center tap to each of the lead ends and infinity to the spark plug leads. [...] Those readings are in the ball park; the primary windings should measure about 0.7 ohms for each section. See http://endwrench.com/images/pdfs/DirectIgnition.pdf for further info. However, in the "Ignition Coil Testing" section of that PDF, the terminal numbers listed are reversed between the primary and secondary (although the diagram is labeled correctly). The correct numbering relative to expected resistance is in http://endwrench.com/images/pdfs/IgnitionCoil.pdf . You obviously recognize the limitations of testing for HV leakage with the low voltage an ohmmeter operates at. A further consideration is that even a single shorted turn in a coil can "kill" the impedance, but often makes almost no difference in the DC resistance, so measuring coil resistance often won't reveal subtle problems. Besides just replacing the coil, a valid test is to use a "ringing" circuit, but that may be beyond how far you're willing to go. Here's a warning about avoiding damage to the coil: http://endwrench.com/images/pdfs/IgnitionCoilSum04.pdf Link to comment Share on other sites More sharing options...
86BRATMAN Posted June 7, 2007 Share Posted June 7, 2007 The timing chain was replaced a few months before it's demise. Not exactly the issue at hand but these cars have a timing belt. Link to comment Share on other sites More sharing options...
DTL Posted June 7, 2007 Author Share Posted June 7, 2007 Ok, Thanks. Y'know, when I put in the first xistor, I thought I smelled the smell of burning electronics pretty soon after I turned the key. I cranked it for maybe 30 sec before I gave up. I am now thinking it will be a bad ECU. If I use a DMM to read the ignitor control signals i was hoping I could determine quickly if it is bad by dc averaging somehow. I not sure if my other meter can do that. Can I put a parallel RC with a bit longer time constant than the expected pulse, ground one side and see what each output charges charges it up to after a few cranks.... (What pulse width should I expect? ) or discharges to... does it goes low ? geeze, a bad/intermittant/floating connector could do that. seems like it would be a bad design if it goes low to fire a spark. if ECU lost power for a bit it would fry the ignitor...I would HATE to fry another ignitor....can a bad key switch do that for me? too many things to think about ;-] Thanks for all the help. Link to comment Share on other sites More sharing options...
DTL Posted June 7, 2007 Author Share Posted June 7, 2007 was replaced ... thats what I thought until I remembered it it died a few months before it really died, and my wife had the work done. Link to comment Share on other sites More sharing options...
hankosolder2 Posted June 7, 2007 Share Posted June 7, 2007 Well, I think the best way to check the ECU signal will be to use that digital 'scope. I'm sure that if you see ANY pulsing on both the drive signals to the ignitor, the ECU is probably fine. What are the odds of it failing in a manner which would allow excessive duty cycle pulses? If you have a Fluke digital multimeter with the analog bargraph display on the bottom of the LCD, you may be able to see some "flickering" as the digital signals pulse at cranking speeds. At the least, you could compare both outputs from the ECU and the DC average should be close if the pulsewidth is similar. I looked at the schematic, and IF the depiction of the ignitor is accurate, there are PNP transistors in it, so yes, a logical low to the ignitor would turn the transistors on and send current to the primary of the coil, so yeah, an open or grounded drive signal could wreak all kinds of havoc. I'm a little suspicious that the internal ignitor schematic isn't accurate (for example, there's no way the drive transistors could withstand the back EMF from the coil pack without clamping diodes in either the coil pack or the ignitor.) So, it very well could use NPN transistors wired with the collectors going to the coil pack and work on logical highs... Chances are that the ECU power lead and the power source for the coil are the same circuit...so if the ECU lost power it wouldn't cook the ignitor. Here's a thought. Why not do something to limit the current going to the coil during testing to protect the ignitor? You could splice some low resistance power resistors in series with each leg of the primary circuit of the coil. (Kind of like "ballast" resistors in old points type ignition.) It would obviously reduce the spark output voltage, but you could try it temporarily during diagnosis and then once you proved that everything was working, wire the coil back in straight. Of course, if the secondary is arcing to the primary, those resistors aren't going to buy you much safety! Good luck, Nathan Ok, Thanks. Y'know, when I put in the first xistor, I thought I smelled the smell of burning electronics pretty soon after I turned the key. I cranked it for maybe 30 sec before I gave up. I am now thinking it will be a bad ECU. If I use a DMM to read the ignitor control signals i was hoping I could determine quickly if it is bad by dc averaging somehow. I not sure if my other meter can do that. Can I put a parallel RC with a bit longer time constant than the expected pulse, ground one side and see what each output charges charges it up to after a few cranks.... (What pulse width should I expect? ) or discharges to... does it goes low ? geeze, a bad/intermittant/floating connector could do that. seems like it would be a bad design if it goes low to fire a spark. if ECU lost power for a bit it would fry the ignitor...I would HATE to fry another ignitor....can a bad key switch do that for me? too many things to think about ;-] Thanks for all the help. Link to comment Share on other sites More sharing options...
The Dude Posted June 7, 2007 Share Posted June 7, 2007 Don't ignition coils usually fire on a collapsing coil field? In others words, current is always running through the primary UNTIL the ECU calls for an ignition spark. Then the current is turned off, the field in primary coil collapses and induces high voltage in the secondary coil. Since the primary is only 12VDC the secondary is a step up coil with many, many windings. Since the ignitor transistor is in the primary circuit no flyback diode is needed. Any voltage re-induced into the primary coil from the secondary would be stepped down. Link to comment Share on other sites More sharing options...
OB99W Posted June 7, 2007 Share Posted June 7, 2007 Don't ignition coils usually fire on a collapsing coil field? In others words, current is always running through the primary UNTIL the ECU calls for an ignition spark. Then the current is turned off, the field in primary coil collapses and induces high voltage in the secondary coil.Yes, sort of; here's some historical perspective: In the Kettering ignition system, the points are referred to as "breaker" (as opposed to "maker") specifically because they interrupt the current flow to the coil primary. (The period during which the primary is connected to the source is called "dwell".) Since the primary is only 12VDC the secondary is a step up coil with many, many windings. Since the ignitor transistor is in the primary circuit no flyback diode is needed.A magnetic field forms when the primary is connected to battery voltage. It collapses when the current is interrupted (by good old breaker points or modern semiconductor); when that happens, due to the inductance of the primary a high voltage spike ("kickback" or "flyback") is produced. See http://www.allaboutcircuits.com/vol_3/chpt_3/8.html for details. Ignition coils do have a high primary/secondary turns ratio, but not as high as you might think, since the flyback voltage is what gets transformed, not just the battery voltage. That flyback voltage is seen by the switching transistors in the igniter, and they need to be rated to withstand it. Using diodes to suppress the kickback defeats the design. Any voltage re-induced into the primary coil from the secondary would be stepped down.True, in terms of the turns ratio. Of course, if there were secondary-to-primary leakage, the voltage wouldn't be limited. Link to comment Share on other sites More sharing options...
DTL Posted June 7, 2007 Author Share Posted June 7, 2007 Ah, another person with an electronics backround. Yes, the plugs paired on one bank of the coil get opposite polarity pulses. It's called a wasted spark system. It seems remotely possible that if you have a open in your ignition secondary circuit (most likely a plug wire) the high voltage could arc internally from the coil pack's secondary winding to the primary winding and fry the transistor in the ignitor. You could check the resistance of the plug wires. I would scrutinize the low voltage part of the circuit first though. Does the coil have equal resistance from the centertap on the primary to each primary leg? Another possiblity might be if the computer is holding one of the outputs to the ignitor constantly low or high...rather than pulsing. The ignitor is probably not designed to dump 100% duty cycle DC into the relatively low resistance of the coil pack. You could check for the proper drive signals going to the ignitor, check the wiring between the ignitor and the coil pack for shorts, check the plug wires for opens, replace the coil pack and ignitor simultaneously and you'll probably be in business. Let us know what you find! Nathan Alrighty then, Dug out my scope from its comf niche and found one output pulsing low for roughly 50 ms every 200ms or so, the other output stayed high and a bit noisier. There are a few places on the web that sells the ECU's. This one: http://www.ecudirect.com/finditfast/finalcombo_yearmakemodel.asp?vmake=SUBARU&vmodel=LEGACY&vyear=1991&type1= sell new ($500) as well as remanufactured($300). But neither they or a couple other companies have exactly my ECU: #31711 AA911. Soooooooooooo What to do? I can get mine repaired I think.... Not sure that the turn around time will be all that quick. I was wondering how important a few digits would be for my model Which is : 1991 Legacy LS 4wd ABS the VIN is 4S4BJ65C2M7900094. I think the car is worth maybe 1k-$1200 here on long island where cars are in high demand. It ran great until it died. $300 more would be about the max I would want to put in it... I would prefer to get away cheaper than that. Any of you have any suggestions for me at this point? What are the limitations as far as using different numbered ECUs? I gueass they are programmed to the specs of the of the sensors and transducers used in this specific model. I was wondering if another LS Legacy's ECU might work well enough??? or can another model be flashed/programmed to work with mine? I took the cover off the ECU, I was wodering how hard it would be to isolate the bad chip/component(s) and get replacement parts. I have done a little surface mount soldering, but not for years, it would give me an excuse to buy a decent surface mount soldering iron on ebay! Hankosolder? Thanks everyone Bill Link to comment Share on other sites More sharing options...
ericem Posted June 7, 2007 Share Posted June 7, 2007 Why not find one used on ebay? or just go to a junkyard? Link to comment Share on other sites More sharing options...
hankosolder2 Posted June 7, 2007 Share Posted June 7, 2007 > That flyback voltage is seen by the switching transistors in the igniter, and >they need to be rated to withstand it. Using diodes to suppress the kickback >defeats the design. Man, oh man, I was wrong about clamping diodes. I guess it's just force of habit-- driving an inductive load= need to clamp back EMF. Or not in this case. In a kettering ignition, the coil forms a resonant circuit with the condenser across the points, right? I always thought that the cap across the points was just to extend the life of the points, but it also really increases the voltage output by tuning the primary circuit to resonate. Food for thought. >True, in terms of the turns ratio. Of course, if there were >secondary-to-primary leakage, the voltage wouldn't be limited. The back EMF still a decent spike, as anyone who has manually opened the points of a kettering ignition with the engine off and key on can attest to! Nathan Link to comment Share on other sites More sharing options...
OB99W Posted June 7, 2007 Share Posted June 7, 2007 In a kettering ignition, the coil forms a resonant circuit with the condenser across the points, right? I always thought that the cap across the points was just to extend the life of the points, but it also really increases the voltage output by tuning the primary circuit to resonate.[...]Exactly. Anyone who's had a bad condenser (what we now call a capacitor) in a breaker-points-operated system knows that it does more than suppress arcing and metal transfer. The resonance causes there to be a damped sinusoid response, which means the plug voltage is not just a "spike", but "stretched", providing the likelihood of more-complete ignition. The back EMF still a decent spike, as anyone who has manually opened the points of a kettering ignition with the engine off and key on can attest to!Even measuring the resistance of any decently-sized coil/solenoid/etc. while holding ohmmeter probe tips can lead to that "shocking" experience. Link to comment Share on other sites More sharing options...
OB99W Posted June 7, 2007 Share Posted June 7, 2007 Okay, I've probably bored some of you enough already, so for those who want further info (waveforms, etc.), I found this: http://www.motor.com/MAGAZINE/Pdf/052005_04.pdf Link to comment Share on other sites More sharing options...
DTL Posted June 8, 2007 Author Share Posted June 8, 2007 OK, I am buying the ECM from an outfit called strangely enough ECM, they also have an airbag business: http://ecmoutlet.com/index.html another $235.00 , but these guys seem to know what they are talking about. They say The software for the ECM is dependendent on the Vehicle VIN and needs to be reflashed. My guess it that the hardware is somewhat generic and the ECU # specifies the flash version. I got a quote for $55.00 from an outfit in Littleton Colorado but, after checking the BBB with there phone # turns out they are not the most reputable group that ever put up a shingle on the internet or anywhere. http://search.bbb.org/Results.aspx?ph=800-944-3215 The other guys were not nearly as bad in there BBB report. I confronted them with the BBB info and the guy said basically that they do hundreds of orders and there are always problems, many times it is customers who don't know what they are doing or, are less than reputable themselves. http://search.detroit.bbb.org/nis/newsearch2.asp?ID=1&ComID=0332000053001171 Then I get an email from the Littleton CO outfit, to whom I sent a link to their BBB record. The guy says that it is under new ownership as of May of this year.... He sounded even a bit embarrassed..... so maybe these guys are worth giving a chance. I went with the $235 ECM anyway, I just don't have time to screw around anymore. I will post again when I install it... hopefully it will all work. Bill Link to comment Share on other sites More sharing options...
Kilroy Posted June 10, 2007 Share Posted June 10, 2007 I paid $52 including shipping for an ECU for my '90 Legacy Wagon Auto. I hope you used: http://www.car-part.com I just did a search and there are 9 pages, you just need to check and be sure the part number matches. My car needed the ECU that was only available for 6 months, had no trouble finding it on this website. Link to comment Share on other sites More sharing options...
WoodsWagon Posted June 10, 2007 Share Posted June 10, 2007 Why not find one used on ebay? or just go to a junkyard? I second this. Way cheaper, and on an older car like yours, you should be familiar with getting parts from a scrapyard. Link to comment Share on other sites More sharing options...
Cougar Posted June 10, 2007 Share Posted June 10, 2007 Okay, I've probably bored some of you enough already, so for those who want further info (waveforms, etc.), I found this: Keep boring us OB99W. This is good info even if it is a refresher. Thanks! Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now